Re: Status of the SPARQL algebra

On Nov 20, 2006, at 10:10 AM, Seaborne, Andy wrote:

>
> I've checked in my working copy of a proposed SPARQL algebra.
>
> http://www.w3.org/2001/sw/DataAccess/rq23/rq24-algebra.html

Awesome.

> 1/ Comments on the structure

I like the structure; FWIW, I would v. much prefer that whatever  
structure is worked out here be preserved w/out interpolation when  
you move this into rq24, which argues for the algebra having an  
entire section of its own. That plus rq24 as it is, w/ the definition  
bits removed, strikes me as a spec v. much worth having (modulo, of  
course, what substantive changes the new algebra implies for the  
other rq24 bits: grammar, exposition, etc.)

Comments:

1. Numbering the sections and subsections will make it easier to  
refer to (yes, I know this makes moving it into rq24 a matter of some  
duplication...)

2. For the Mapping a SPARQL Query to the SPARQL Algebra section, it's  
not v. clear what Step 1, 2, and 3 are... I assume or guess they are  
steps to transform an AST into something, a set or tree of algebraic  
expressions? The rule in Step 1 could be clearer; the "it consists  
of" locution should be replaced by some term... It's just sorta  
ambiguous as-is.

3. FWIW, I very much like this style of examples where you show  
surface syntax transformed into algebraic expressions. It should be  
v. helpful to implementers and I think it satisfies one of Fred's  
points about the spec not clearly enough tying grammar to query  
evaluation.

Good show, Andy, even though this is preliminary and a bit skeletal,  
it's a huge improvement in spec style, for my money.

> 4/ Advice on how to typeset squiggly symbols in portable XHTML,

FWIW, on F'fox 2.0 on Mac, this looks fine.

Cheers,
Kendall

Received on Monday, 20 November 2006 16:44:20 UTC